Techniques to organize test results

• US 7,333,962 B2
• Filed: 02/22/2006
• Issued: 02/19/2008
• Est. Priority Date: 02/22/2006
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Zen, Lee-Ming
• Primary Examiner(s): STARKS, WILBERT L
• Application Number:
  US11/359,742
• Publication Number:
  US 20070198445A1
• Time in Patent Office: 727 Days
• Field of Search: 706/15, 706/45, 382/132, 436/43, 600/450, 700/17, 700/83, 713/600
• US Class Current: 706/15
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

Techniques to organize test results are described. In one embodiment, for example, vectors with failure information may be generated from test result files. Nodes for a self-organizing map may be trained with multiple vectors. Vector groups may be formed for multiple nodes with the self-organizing map. Other embodiments are described and claimed.

Claims

1. A method, comprising:

generating vectors with failure information from test result files;

training nodes for a self-organizing map with multiple vectors; and

forming vector groups for multiple nodes with said self-organizing map.

2. The method of claim 1, comprising displaying said vector groups.

3. The method of claim 1, comprising:

extracting a string and an image from a test result file;

generating an image histogram for said image; and

generating a vector for said test result file with said string and said image histogram.

4. The method of claim 1, comprising:

initializing a node string value for a node;

comparing a vector string value for a vector with said node string value;

generating a string distance value between said vector string value and said node string value; and

modifying said node string value using said string distance value.

5. The method of claim 1, comprising:

comparing a vector string value for a vector with a node string value for a node;

generating a string distance value between said vector string value and said node string value; and

mapping said vector to said node using said string distance value.

6. The method of claim 1, comprising:

initializing a node image value for a node;

comparing a vector image value for a vector with said node image value;

generating an image distance value between said vector image value and said node image value; and

modifying said node image value using said image distance value.

7. The method of claim 1, comprising:

comparing a vector image value for a vector with a node image value for a node;

generating an image distance value between said vector image value and said node image value; and

mapping said vector to said node using said image distance value.

8. An article comprising a machine-readable storage medium containing instructions that if executed enable a system to generate vectors with failure information from test result files, train nodes for a self-organizing map with multiple vectors, and form vector groups for multiple nodes with said self-organizing map.

9. The article of claim 8, comprising instructions that if executed enable the system to display said vector groups.

10. The article of claim 8, comprising instructions that if executed enable the system to extract a string and an image from a test result file, generate an image histogram for said image, and generate a vector for said test result file with said string and said image histogram.

11. The article of claim 8, comprising instructions that if executed enable the system to initialize a node string value for a node, compare a vector string value for a vector with said node string value, generate a string distance value between said vector string value and said node string value, and modify said node string value using said string distance value.

12. The article of claim 8, comprising instructions that if executed enable the system to compare a vector string value for a vector with a node string value for a node, generate a string distance value between said vector string value and said node string value, and map said vector to said node using said string distance value.

13. The article of claim 8, comprising instructions that if executed enable the system to initialize a node image value for a node, compare a vector image value for a vector with said node image value, generate an image distance value between said vector image value and said node image value, and modify said node image value using said image distance value.

14. The article of claim 8, comprising instructions that if executed enable the system to compare a vector image value for a vector with a node image value for a node, generate an image distance value between said vector image value and said node image value, and map said vector to said node using said image distance value.

15. The article of claim 8, said machine-readable storage medium comprising a processing device, a computer-readable medium, a communications medium, or a propagated signal.

16. An apparatus, comprising:

a vector generator to generate vectors with failure information from test result files;

a trainer to train nodes of a self-organizing map with multiple vectors; and

a classifier to form vector groups for multiple nodes with said self-organizing map.

17. The apparatus of claim 16, comprising a displayer to display said vector groups using a graphics user interface.

18. The apparatus of claim 16, with each vector to comprise a string and an image histogram.

19. The apparatus of claim 16, said self-organizing map to comprise multiple nodes connected by node links.

20. The apparatus of claim 16, comprising:

a testing server array comprising multiple testing servers to provide test instructions; and

a testing client array comprising multiple testing units to output test result files in response to said test instructions.